Abstract:Partial differential equations (PDEs) play a fundamental role in modeling and simulating problems across a wide range of disciplines. Recent advances in deep learning have shown the great potential of physics-informed neural networks (PINNs) to solve PDEs as a basis for data-driven modeling and inverse analysis. However, the majority of existing PINN methods, based on fully-connected NNs, pose intrinsic limitations to low-dimensional spatiotemporal parameterizations. Moreover, since the initial/boundary conditions (I/BCs) are softly imposed via penalty, the solution quality heavily relies on hyperparameter tuning. To this end, we propose the novel physics-informed convolutional-recurrent learning architectures (PhyCRNet and PhyCRNet-s) for solving PDEs without any labeled data. Specifically, an encoder–decoder convolutional long short-term memory network is proposed for low-dimensional spatial feature extraction and temporal evolution learning. The loss function is defined as the aggregated discretized PDE residuals, while the I/BCs are hard-encoded in the network to ensure forcible satisfaction (e.g., periodic boundary padding). The networks are further enhanced by autoregressive and residual connections that explicitly simulate time marching. The performance of our proposed methods has been assessed by solving three nonlinear PDEs (e.g., 2D Burgers’ equations, the λ - ω and FitzHugh Nagumo reaction–diffusion equations), and compared against the start-of-the-art baseline algorithms. The numerical results demonstrate the superiority of our proposed methodology in the context of solution accuracy, extrapolability and generalizability.

Neural PDE Solvers for Irregular Domains

PhyGeoNet: Physics-informed Geometry-Adaptive Convolutional Neural Networks for Solving Parameterized Steady-State PDEs on Irregular Domain

Non-equispaced Fourier Neural Solvers for PDEs

Learning Only On Boundaries: a Physics-Informed Neural operator for Solving Parametric Partial Differential Equations in Complex Geometries

Solving inverse-PDE problems with physics-aware neural networks

Solving Irregular and Data-enriched Differential Equations using Deep Neural Networks

Positional Embeddings for Solving PDEs with Evolutional Deep Neural Networks

Solving partial differential equations with sampled neural networks

Neural Q-learning for solving PDEs

Inf-Sup neural networks for high-dimensional elliptic PDE problems

Message Passing Neural PDE Solvers

Numerical solutions of boundary problems in partial differential equations: A deep learning framework with Green's function

Neural-PDE: a RNN Based Neural Network for Solving Time Dependent PDEs

Solving Partial Differential Equations Using Deep Learning and Physical Constraints

Domain Adaptation Based Transfer Learning Approach for Solving PDEs on Complex Geometries

Solving PDEs on unknown manifolds with machine learning

Graph Neural PDE Solvers with Conservation and Similarity-Equivariance

PhyCRNet: Physics-informed convolutional-recurrent network for solving spatiotemporal PDEs